When geometry is not enough, Riemann sums approximate the definite integral by summing products of function values and subinterval widths. Left, right, midpoint, and trapezoidal sums each have predictable over- or underestimate behavior based on whether the function is increasing, decreasing, concave up, or concave down. As the maximum subinterval width approaches zero, the Riemann sum limit equals the definite integral, written as the integral from a to b of f(x) dx.
- Left Riemann sum: Uses the left endpoint of each subinterval as the rectangle height; overestimates when f is decreasing, underestimates when increasing.
- Right Riemann sum: Uses the right endpoint; overestimates when f is increasing, underestimates when decreasing.
- Trapezoidal sum: Averages left and right endpoint values; overestimates when f is concave up, underestimates when concave down.
- Definite integral as limit: The integral from a to b of f(x) dx equals the limit of the Riemann sum as the maximum subinterval width approaches zero.
- Summation notation: A Riemann sum is written as the sum from i=1 to n of f(x_i*) times delta x_i, where x_i* is a sample point in the ith subinterval.
Given a table of values with unequal subintervals, can you compute a left, right, and trapezoidal Riemann sum and identify which is an overestimate?
| Method | Height used | Overestimates when | Underestimates when |
|---|
| Left sum | Left endpoint | f is decreasing | f is increasing |
| Right sum | Right endpoint | f is increasing | f is decreasing |
| Midpoint sum | Midpoint of subinterval | f is concave down | f is concave up |
| Trapezoidal sum | Average of endpoints | f is concave up | f is concave down |